Key Variables Sample Clauses

Key Variables. The dependent variable in all of my regression models is a form of trading volume. A high level of trading volume is indicative of a liquid market. I specifically examine turnover. I calculate abnormal turnover according to the model of ▇▇▇▇ (1999) which argues that individual firm turnover should equal aggregate market turnover under normal circumstances. That is, trading in individual firms should be similar to the level of trading in the aggregate market. Any deviation from this standard would be considered ‘abnormal’. This argument is based on the theoretical prediction that market- wide trading translates into trading in each asset according to its relative value in the market11. In the context of informed versus uninformed traders, this model separates the informed from the uninformed traders so that an analysis of the resulting abnormal (idiosyncratic) volume actually allows me to examine the behavior of informed traders. Therefore, I measure turnover (Turnover) for an individual firm as shares traded shares outstanding (1) 10 While this methodology results in a loss of approximately 82% of the original sample, this reducation is comparable to that of Barron (1995) who focuses on a reduced sample of 1,520 observations from a full sample of 6,727 observations – a reduction of 77%. 11 According to Tkac, although this model is simple, it should isolate idiosyncratic trading activity. Further, Tkac attributes undertrading (overtrading) possibly to less (more) non-rebalancing activity. In other words, fewer (more) investors are trading these stocks based on firm-specific information. I measure market turnover using data from all firms (i) in the CRSP database. I calculate market turnover (MktTurnover) as ∑(shares traded)i ∑( ) MktTurnover = i (2) I calculate abnormal turnover, then, as AbnormalTO = Turnover − MktTurnover (3) I use several explanatory variables related to belief dispersion and the characteristics of belief revisions. With respect to belief dispersion, I first address overall dispersion followed by dispersion within groups of investors. These groups, ‘holders’ and ‘non-holders’, represent investors that own or do not own the traded asset, respectively. I or HDispjt = σ forecasts jt >μ (4) forecasts jt >μ NHDispjt = σ forecasts jt <μ (5) forecasts jt <μ where HDispjt represents dispersion within the ‘holder’ group and is measured as coefficient of variation in the forecasts greater than (less than) the mean value of all forecasts for the sa...